Skip to main content
Fig. 1 | Journal of Translational Medicine

Fig. 1

From: Rapid diagnosis and comprehensive bacteria profiling of sepsis based on cell-free DNA

Fig. 1

An illustration of our approach to sepsis diagnosis and bacteria inference based on cell-free DNA (cfDNA). a We used two public cfDNA datasets to obtain 38 sepsis and 118 healthy samples. All human reads were removed from the datasets using Bowtie2. Through alignment and classification, the normalized abundances of bacteria were estimated from the remaining non-human reads using Centrifuge [27]. b Our diagnosis strategy is a two-step procedure based solely on cfDNA from blood. First, we selected candidate pathogenic bacterial species through statistical analysis (see “Methods”). Second, a Random Forest is used to calculate a diagnosis score for each sample. c Due to the limited volume of a blood sample, not all bacterial species will be identified in cfDNA sequencing data. Using the bacterial co-occurrence network, we developed a method to infer unobserved bacterial species

Back to article page